The more useful AI turns out to be, the more electricity we will need to power it.
Forget the worries about Skynet. Could AI be evolving into something even worse: a climate villain?
In an article for The New Yorker (March 9), Elizabeth Kolbert discusses the vast amounts of energy consumed by Bitcoin, before turning to a new electroglutton, AI:
In a paper he published last fall, in Joule, a journal devoted to sustainable energy,[Alex] de Vries, who now works for the Netherlands’ central bank, estimated that if Google were to integrate generative A.I. into every search, its electricity use would rise to something like twenty-nine billion kilowatt-hours per year. This is more than is consumed by many countries, including Kenya, Guatemala, and Croatia.
And from power consumption, Kolbert moves grimly to climate change:
Last week, the International Energy Agency announced that energy-related global CO2 emissions rose, yet again, in 2023, to more than thirty-seven billion metric tons. The increase comes at a time when the whole world is supposedly striving to reach net-zero emissions, and it indicates that global efforts are, to put it mildly, falling short. Much of the increase in emissions came from China, and most of it was driven by century-old technologies, such as the internal-combustion engine. So data centers are, for now at least, a small part of the problem. Still, as the use of A.I. ramps up and bitcoin prices reach new heights, the question is: How can the world reach net zero if it keeps inventing new ways to consume energy? (In the U.S., data centers now account for about four per cent of electricity consumption, and that figure is expected to climb to six per cent by 2026.)
To ask the question how the world can reach net zero “if it keeps inventing new ways to consume energy” is (in some respects) to ask how the world can reach net zero if humanity continues to innovate. Obviously, not every innovation will involve increased energy consumption, but many will. Even innovations that increase energy efficiency may ultimately boost energy consumption, a paradox named after the 19th century English economist William Jevons. Jevons argued that it was “wholly a confusion of ideas to suppose that the economical use of fuel is equivalent to a diminished consumption. The very contrary is the truth.” The reduction in the cost of energy would eventually encourage an increase in its use that, in absolute terms, would more than cancel out any fall in demand attributable to improved efficiency.
On the topic of AI, Kolbert explains that:
The kind of machine learning that produced ChatGPT relies on models that process fantastic amounts of information, and every bit of processing takes energy. When ChatGPT spits out information (or writes someone’s high-school essay), that, too, requires a lot of processing. It’s been estimated that ChatGPT is responding to something like two hundred million requests per day, and, in so doing, is consuming more than half a million kilowatt-hours of electricity. (For comparison’s sake, the average U.S. household consumes twenty-nine kilowatt-hours a day.)
To be fair, Kolbert notes that:
A.I. could potentially be used to alleviate some of the problems it is exacerbating. For instance, it might be used to improve the efficiency of renewable-energy systems, which could reduce emissions from server farms. But it seems unlikely that such gains will keep up with A.I.’s growing electricity demands.
The truth of that observation is reinforced by the Jevons paradox. If AI helps us devise more efficient uses of electricity, we will use more electricity.
Kolbert also quotes OpenAI’s Sam Altman:
“There’s a fundamental mismatch between this technology [A.I.] and environmental sustainability,” de Vries said. Recently, the world’s most prominent A.I. cheerleader, Sam Altman, the C.E.O. of OpenAI, voiced similar concerns, albeit with a different spin. “I think we still don’t appreciate the energy needs of this technology,” Altman said at a public appearance in Davos. He didn’t see how these needs could be met, he went on, “without a breakthrough.” He added, “We need fusion or we need, like, radically cheaper solar plus storage, or something, at massive scale — like, a scale that no one is really planning for.”
We’ll see, but if Altman is right, such a surge in the demand for electricity would merely reflect AI’s value. The more useful AI turns out to be, the more electricity we will need to power it. Could accommodating that demand clash with the (almost certainly) already doomed objective of achieving net zero by 2050? Yes.
But even if AI is not the trigger, sooner or later a choice will have to be made between (1) preserving the remarkable two-centuries-long growth surge that has brought (and is bringing) so much prosperity to so many or (2) flattening its trajectory in order to hit net zero. Those in the poorer parts of the world have shown little interest in accepting any slowdown in the improvement in their standards of living in exchange for a reduction in greenhouse-gas emissions. And, with noisy exceptions, the same will also be true for those living in wealthier parts of the planet — whatever they may tell pollsters now. That means that the inevitable rationing, restrictions and economic slowing imposed in the name of carbon-paring will almost certainly lead to social disorder, authoritarianism or, quite probably, both.
We should thus keep aiming to maximize economic growth. If AI can help in that objective, so much the better. Among its other benefits, the wealth created will help fund the technologies and pay for the resilience that will enable both the rich and poor to mitigate a changing climate and cope with its effects.